Mice are widely used in genetic and medical research. Because genetic alterations and exposure to drugs can induce subtle behavioral changes, there is a need to automatically recognize and quantify a variety of murine behaviors. We propose a system for automated behavioral phenotyping of mice. The system will acquire digital video data of single or multiple mice on typical bedding material, under day and night lighting conditions. The system will find the mouse contour, the location and orientation of its head, and the tail position. With this information, we can categorize a variety of motions, including locomotion, freezing, grooming, rearing, seizures, stereotypies, and a variety of social behaviors including nosing, ano-genital sniffing, aggression, and mating behaviors. A microphone will be used for recording vocalizations and a speaker will be used for startle tests. Currently, companies screen thousands of mutant mice using devices that are not automated. Thus phenotyping is time consuming, labor intensive, and limited to easily measured physical quantities and common behaviors. Our system will be a commercially viable tool useful to drug companies and scientists, will aid discovery in mouse social behavior, and will help researchers understand how genetic or pharmacological manipulations cause changes in behavior.